translate

Deaf and Hard of Hearing

British Sign Language Corpus

A. Schembri, J. Fenlon, R. Rentelis, S. Reynolds, K. Cormier

Data from deaf native, near-native and fluent signers across the United Kingdom were collected to create a machine-readable digital corpus of spontaneous and elicited British Sign Language (BSL) digital video. Participants were filmed in pairs with one camera focused on each participant and a third focused on both members of the pair.
Tags: 
Sign languageBritish Sign Language

SLANG3k

P. Toman

The study investigates automatic identification of internet-based signed language data using Neural Networks. Online videos were searched for different sign languages.
Tags: 
Sign languageAmerican Sign LanguageBritish Sign LanguageGerman sign language

RWTH-PHOENIX-Weather 2014 T

N. Camgoz, S. Hadfield, O. Koller, H. Ney, R. Bowden

This work aims to generate spoken language translations from sign language videos, taking into account the different word orders and grammar, and argues that existing linguistic datasets are too small to support deep learning. The study uses the daily news and weather forecast airings of the German public tv-station featuring sign language interpretation. The corpus covers a vocabulary of 1066 different signs annotated using sign glosses by deaf specialists
Tags: 
Sign languageGerman sign languageGerman

Swedish Sign Language Dictionary

J. Mesch, L. Wallin, T. Björkstr

This dictionary includes video files for signs in Swedish Sign Language created for general educational resources. Each sign entry on the web-based dictionary consists of sign demonstration, variants by one or two hands, usage examples (sentence), and photo and sign notation.
Tags: 
Sign languageSwedish sign languageSwedish

The ASL Lexicon Video Dataset

V. Athitsos, C. Neidle, S. Sclaroff, J. Nash, A. Stefan, Q. Yuan, A. Thangali

Native ASL signers perform different signs as part of a project to develop a computer vision system that allows users to look up the meaning of an ASL sign. The video sequences are collected in a recording facility, using a four-camera system that simultaneously captures two frontal views, one side view, and one view zoomed in on the face of the signer.
Tags: 
Sign languageAmerican Sign Language

AUTSL: A Large Scale Multi-Modal Turkish Sign Language Dataset

O. Sincan, H. Keles

This work collects a large dataset with different challenging backgrounds that is suitable for modelling a realistic recognition system with real life scenarios for Turkish Sign Language. 226 signs and 38,336 isolated video samples were recorded in 20 different background conditions with Microsoft Kinect v2 containing color image (RGB), depth, and skeleton modalities.
Tags: 
Sign languageTurkish sign languageTurkish

Alignment Data base for a Sign Language Concordancer

M. Kaczmarek, M. Filhol

Translations and interpretation of sign language are distinct. This project aligns the video-signed translation to their interpretations and build a dataset for training an automatic concordancer. RGB cameras recorded the gestures was used to alight the translations to the interpretations of the corpus, using the French sub-set of Dicta-Sign elicitation content.
Tags: 
Sign language

The Korean Sign Language Dataset for Action Recognition

S. Yang, S. Jung, H. Kang, C. Kim

A Korean sign language corpus is collected using RGB video cameras of sign units to train a convolutional neural network for predicting the sign. RGB camera recorded the gestures for 77 words in Korean Sign language.
Tags: 
Sign languageKorean Sign Language

WLASL: A Large-Scale Dataset for Word-Level American Sign Language

D. Li, C. Opazo, X. Yu, H. Li

This dataset is introduced to facilitate word level sign recognition research. Two main sources from the Internet, multiple educational sign language websites, such as ASLU and ASL-LEX and ASL tutorial videos on YouTube are used for data collection. 21,083 videos performed by 119 signers are extracted, and each video only contains one sign in ASL. Each sign is performed by at least 3 different signers for diversity
Tags: 
Sign languageAmerican Sign Language

ChicagoFSWild

B. Shi, A. Rio, J. Keane, J. Michaux, D. Brentari, G. Shakhnarovich, K. Livescu

The dataset is created for recognizing American Sign Language fingerspelling in the wild that is in naturally occurring video. Data consist of manullay located fingerspelling clips from ASL videos on YouTube, aslized.org and deafvideo.tv. with videos annotated by in-house annotators.
Tags: 
Sign languageAmerican Sign Language